Robotic Multi-pass Weld Seam Correction Control Based On Visual Sensing

With the application of industrial robots in the field of welding more and more widely,automation,digitization and intelligence have become the important development direction of welding process.So,it is greatly significant to study the process of weld seam correction control to promote the development of welding industry.In the practical industrial application of multi-pass welding seam tracking for heavy plate,most of the laser seam tracking products on the market can only track the root pass,however,it is difficult to track the subsequent passes.Therefore,the purpose of this paper is to explore a robotic weld seam correction control strategy based on visual sensing and to establish the corresponding visual sensing weld seam correction control system,solving the multi-pass welding seam tracking problem gradually.Based on the previous generation system,the visual sensing weld seam correction system is improved to satisfy the complex actual production conditions.The industrial camera's installation location is changed from the front of the welding torch to the top left of the front and the monitoring interface of the human-machine interaction module is also redesigned.The improved system has a wider range of application,a simpler monitoring interface,a simpler operation process,stronger judgment and recognition ability,and an enhanced anti-arc ability,which is more suitable for filling passes.This paper explores the influence of welding torch offset on weld images which are obtained by the industrial camera placed on the side of welding torch,and analyzes and summarizes the relationship between them.Then,the feature points are extracted to judge whether the welding torch deviates from target.A corresponding judgment and correction control strategy is proposed.Then,the feasibility of the correction control strategy is verified with the monitoring interface.In addition,this paper also proposes a matrix mathematical model which can automatically identify and judge the welding torch's offset,which contributes to provide the theoretical possibility for fully automatic seam tracking.Base on the multi-pass welding planning scheme for symmetric V-groove heavy plate,the accuracy and stability of the weld seam correction control system are tested.When the robot is running on the designed multi-pass welding path,the operators can judge whether the welding torch needs to be adjusted according to the real-time welding images from monitoring interface and the correction control strategy,and send a correction command to the robot to achieve the correction control process based on the human-machine interaction mode.By comparing the deviation correction of robot with the actual deviation of weld seam,the accuracy of system is rated quantitatively,which ranges in 0.5 mm and is acceptable.By comparing the actual weld height and width after correction and the designed height and width,the stability of system is tested qualitatively,which ranges in 1 mm and is also acceptable.Moreover,when the system works,the influence of strong arc light can be overcome,the obtained images are clear,the images can be transmitted smoothly,the response rate is sensitive,the correction control process is relatively smooth,no apparent adjustment trace is shown,and the system can track the weld in real time.